Most of the focus of the new Colliergate release is on the new and improved parser. The parser claims up to 24% increases in performance, and that's because it finally brings a new means of peeking.
What I'm talking about is that the Zenith Parsing Engine version 1.5 no longer wastes parsing time when peeking - once a token has been read, it is stored and it's never reparsed. Because of this, when ZPE asks the parser to peek ahead, the parser checks if the token has been parsed, and if not, it parses it, then returns that token. See, whenever you do a K(n) peek ahead, you parse n tokens, but in previous versions of the Zenith Parsing Engine, these tokens were discarded and reparsed again when the parser re-peeked at them or tokenised them, thus, storing a token cache, we can easily discern its value without reparsing it.
It's just astoundingly fast, and this improves CSV, XML, JSON, TOML, JBML and INI parsing too.
What this means is that ZPE is catching up to competitors for parsing speed, and there's still room for further enhancements that are planned for version 1.6.
